MoonshotAI / MoonshotAI/kimi-cli
[Feature Request] --yolo should not auto-approve plans in Plan Mode; separate tool approval from plan review
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 11.4k
- Forks
- 1.3k
- Avg merge
- 9h 47m
- Merged PRs (30d)
- 2
Description
What feature would you like to see?
Currently, when --yolo (or --yes / --auto-approve) is enabled, entering Plan Mode causes the generated plan to be auto-approved and executed immediately without user confirmation.
This behavior conflates two very different levels of risk:
| Level | Risk | User Intent |
|---|---|---|
| Plan review | Low — just reviewing the strategy | Even users who trust the AI may want to verify the direction before execution |
| Tool execution | High — writing files, running shell commands | This is what --yolo is meant to skip |
Expected behavior:
--yoloshould auto-approve tool executions (e.g.,WriteFile,Shell), but it should still pause for plan confirmation in Plan Mode.- Alternatively, provide finer-grained flags such as
--auto-approve-toolsvs.--auto-approve-plans.
Why this matters:
Plan Mode's core value is "plan first, execute second." If the plan itself is automatically approved, Plan Mode loses its primary purpose for users who simply want to avoid clicking "yes" on every file change.
Additional information
Related documentation: https://moonshotai.github.io/kimi-cli/zh/reference/kimi-command.html#审批控制
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by tracing how the --yolo, --yes, and --auto-approve flags are handled alongside the Plan Mode approval flow. Confirm where plan confirmation and tool execution approval are coupled, then verify that plans still pause for review while tools remain auto-approved; the issue does not name specific files or tests.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- cli
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 48/100